Phenotyping Algorithms Research Articles

Identifying patients with neurofibromatosis type 1 related optic pathway glioma using the OMOP CDM.

Identifying patients with neurofibromatosis type 1 related optic pathway glioma using the OMOP CDM.

Leveraging undecided cases in chart-reviewed phenotypes to enhance EHR-based association studies.

Leveraging undecided cases in chart-reviewed phenotypes to enhance EHR-based association studies.

Unmet social needs and diverticulitis: a phenotyping algorithm and cross-sectional analysis.

To validate a phenotyping algorithm for gradations of diverticular disease severity and investigate relationships between unmet social needs and disease severity. An algorithm was designed in the All of Us Research Program to identify diverticulosis, mild diverticulitis, and operative or recurrent diverticulitis requiring multiple inpatient admissions. This was validated in an independent institution and applied to a cohort in the All of Us Research Program. Distributions of individual-level social barriers were compared across quintiles of an area-level index through fold enrichment of the barrier in the fifth (most deprived) quintile relative to the first (least deprived) quintile. Social needs of food insecurity, housing instability, and care access were included in logistic regression to assess association with disease severity. Across disease severity groups, the phenotyping algorithm had positive predictive values ranging from 0.87 to 0.97 and negative predictive values ranging from 0.97 to 0.99. Unmet social needs were variably distributed when comparing the most to the least deprived quintile of the area-level deprivation index (fold enrichment ranging from 0.53 to 15). Relative to a reference of diverticulosis, an unmet social need was associated with greater odds of operative or recurrent inpatient diverticulitis (OR [95% CI] 1.61 [1.19-2.17]). Understanding the landscape of social barriers in disease-specific cohorts may facilitate a targeted approach when addressing these needs in clinical settings. Using a validated phenotyping algorithm for diverticular disease severity, unmet social needs were found to be associated with greater severity of diverticulitis presentation.

Automated identification of clinically meaningful biomechanical phenotypes in cerebral palsy through multicenter gait data.

Automated identification of clinically meaningful biomechanical phenotypes in cerebral palsy through multicenter gait data.

Leveraging electronic health records to examine differential clinical outcomes in people with Alzheimer's Disease.

Alzheimer's disease (AD) carries a high societal burden inequitably distributed across demographic groups. To examine differences in readily ascertainable clinical outcomes of AD decline among demographic groups. Leveraging the electronic health record (EHR) data (1994-2022) of patients with ≥1 diagnosis code for AD or related dementia from two large healthcare systems, we applied a knowledge graph-guided unsupervised phenotyping algorithm to predict AD diagnosis status and validated using gold-standard chart-reviewed and registry-derived diagnosis labels. After excluding patients with <24 months of data or who were admitted to nursing homes prior to AD diagnosis, we performed survival analyses at each healthcare system to assess the time to two readily ascertainable clinical outcomes of AD decline ( i.e., nursing home admission, death), stratified by demographic groups and accounting for baseline covariates ( e.g., age, gender, race, ethnicity, healthcare utilization, and comorbidities). We then performed a fixed-effects meta-analysis of the survival analysis data from both healthcare systems. The AD diagnosis phenotyping algorithm demonstrated high accuracy in identifying AD patients across both healthcare systems (AUROC score range: 0.835-0.923). Of the 34,181 AD patients in both healthcare systems (62% women, 90% non-Hispanic White, 80.39±9.28 years of age at AD diagnosis), 32% were admitted to nursing homes and 50% died during follow- up. In the fixed-effect meta-analysis, non-Hispanic White patients had a lower risk of nursing home admission (HR[95% CI]=0.825[0.776-0.877], p <0.001) and higher risk of death (HR[95% CI]=1.381[1.283-1.487], p <.0001) than racial and ethnic minorities. There was no difference between women and men in their risk of nursing home admission (HR[95% CI]=1.008[0.967-1.050], p =.762), but women had a lower risk of death (HR[95% CI]=0.873[0.837-0.910], p <.0001) than men. This study creates two large EHR-based AD cohorts and adds to the real-world evidence of demographic differences in clinical AD decline, which could potentially inform individual clinical management and future public health policies.

The Relationship between Visceral Fat Accumulation and Risk of Cardiometabolic Multimorbidity: The Roles of Accelerated Biological Aging.

To investigate the association between visceral fat accumulation and the risk of cardiometabolic multimorbidity (CMM) and the potential roles of accelerated biological aging in this relationship. Using data from the UK Biobank, a nationwide cohort study was conducted using the available baseline body roundness index (BRI) measurement. Biological aging was assessed using the Klemera-Doubal method for biological age and the phenotypic age algorithms. The association between the BRI and CMM was estimated using the Cox proportional hazards regression model, while the roles of biological aging were examined through interaction and mediation analyses. During a median follow-up of 14.52 years, 6156 cases of CMM were identified. A significant association was observed between the BRI and CMM. The hazard ratio (HR) for CMM was 3.72 (95% confidence interval [CI]: 3.35-4.13) for individuals in the highest quartile compared with those in the lowest quartile of the BRI. More importantly, the BRI (AUC, 0.701; 95% CI, 0.694-0.707) demonstrated superior predictive performance relative to body mass index (AUC, 0.657; 95% CI, 0.650-0.664). Furthermore, the BRI exhibited additive interactions with accelerated biological aging on the risk of CMM, and accelerated biological aging partially mediated the association between the BRI and CMM. These findings provide evidence for the application of the BRI as a novel and readily accessible screening tool associated with CMM, suggesting that the effective management of visceral fat and biological aging deceleration may hold promise for reducing CMM risk.

Diagnostic algorithm for the detection of carbapenemases and extended-spectrum β-lactamases in carbapenem-resistant Pseudomonas aeruginosa.

Pseudomonas aeruginosa can acquire carbapenem resistance through various mechanisms, including genomic mutations leading to the overexpression of efflux pumps, intrinsic AmpC-β-lactamase, and/or reduced permeability, and/or through the acquisition of plasmid-mediated carbapenemases and/or extended-spectrum-β-lactamases (ESBLs). Unfortunately, carbapenemase/ESBL-producing-carbapenem-resistant-P. aeruginosa (CP/ESBL-CRPA) cannot be differentiated from non-CP/ESBL-CRPA based solely on susceptibility testing results of conventional β-lactam (BL)-antibiotics. Knowing that these two groups display different activity profiles toward novel BL/β-lactamase-inhibitor (BLI) combinations, we developed and verified a cost-effective and easy-to-use diagnostic algorithm for screening and differentiation of carbapenemase and ESBL production in CRPA. We determined disc diffusion inhibition zones and gradient strip minimal inhibitory concentration values of 136 whole-genome-sequenced CRPA (70 metallo-β-lactamase-[MBL-], 1 GES-5-, 1 KPC-2-, 12 ESBL-, and 53 AmpC-hyper-producing isolates). We used the following BL-BLI combinations: ceftolozane-tazobactam (C-T), ceftazidime-avibactam, imipenem-relebactam (I-R), meropenem-vaborbactam, cefepime-enmetazobactam (C-E), and aztreonam-avibactam. We also included a lateral flow immunoassay (Carba-5, NG-Biotech) for confirmation of MBL production and double disc synergy testing (DDST) to improve ESBL detection. C-T was the most effective screening antibiotic for distinguishing MBL and ESBL producers from AmpC-hyperproducing CRPA, achieving a sensitivity of 100% for both MBL and ESBL producers. I-R reliably confirmed MBL production in C-T positive screened CRPAs, with a sensitivity of 92.8% and specificity of 100%. Incorporating Carba-5 into the phenotypic algorithm improved sensitivity for confirming MBL production to 100%. For the remaining C-T positive but I-R negative isolates, C-E showed 75% sensitivity and 78.6% specificity in detecting ESBL production. The DDST further confirmed ESBL production in six out of nine ESBL producers (66.6%). In conclusion, we established a simple and cost-effective diagnostic algorithm, enabling screening and confirmation of carbapenemase and ESBL production in CRPA.IMPORTANCECarbapenem-resistant Pseudomonas aeruginosa (CRPA) is a major global health threat, and rapid identification of its resistance mechanisms is crucial for effective treatment and infection control. Differentiating between carbapenemase-producing (CP), extended-spectrum β-lactamase-producing (ESBL), and AmpC-hyperproducing CRPA is challenging, as conventional susceptibility testing cannot reliably distinguish these resistance mechanisms. Our study presents a simple, cost-effective, and easy-to-implement phenotypic diagnostic algorithm that enables accurate screening and confirmation of CP and ESBL production in CRPA. This method is particularly valuable for laboratories lacking access to molecular diagnostics, as it provides a practical alternative for routine testing. By facilitating the early detection of resistant P. aeruginosa strains, this approach has the potential to improve patient outcomes, optimize antimicrobial therapy, and enhance global surveillance efforts against multidrug-resistant pathogens.

Standardizing phenotypic algorithms for the classification of degenerative rotator cuff tear from electronic health record systems.

Degenerative rotator cuff tears (DCTs) are the leading cause of shoulder pain, affecting 30%-50% of individuals over 50. Current phenotyping strategies for DCT use heterogeneous combinations of procedural and diagnostic codes and are concerning for misclassification. The objective of this study was to create standardized phenotypic algorithms to classify DCT status across electronic health record (EHR) systems. Using a de-identified EHR system, containing chart level data for ∼3.5 million individuals from January 1998 to December 2023, we developed and validated 2 types of algorithms-one requiring and one without imaging verification-to identify DCT cases and controls. The algorithms used combinations of International Classification of Diseases (ICD) / Current Procedural Terminology (CPT) codes and natural language processing (NLP) to increase diagnostic certainty. These hand-crafted algorithms underwent iterative refinement with manual chart review by trained personnel blinded to case-control determinations to compute positive predictive value (PPV) and negative predictive value (NPV). The algorithm development process resulted in 5 algorithms to identify patients with or without DCT with an overall predictive value of 94.5%: (1) code only cases that required imaging confirmation (PPV = 89%), (2) code only cases that did not require imaging verification (PPV = 92%), (3) NLP-based cases that did not require imaging verification (PPV = 89%), (4) code-based controls that required imaging confirmation (NPV = 90%), and (5) code and NLP-based controls that did not require imaging verification (NPV = 100%). External validation demonstrated 94% sensitivity and 75% specificity for the code-only algorithms. This work highlights the inaccuracy of previous approaches to phenotypic assessment of DCT reliant solely on ICD and CPT codes and demonstrate that integrating temporal and frequency requirements, as well as NLP, substantially increases predictive value. However, while the inclusion of imaging verification enhances diagnostic confidence, it also reduces sample size without necessarily improving predictive value, underscoring the need for a balance between precision and scalability in phenotypic definitions for large-scale genetic and clinical research. These algorithms represent an improvement over prior DCT phenotyping strategies and can be useful in large-scale EHR studies.

Knowledge-Driven Online Multimodal Automated Phenotyping System.

Though electronic health record (EHR) systems are a rich repository of clinical information with large potential, the use of EHR-based phenotyping algorithms is often hindered by inaccurate diagnostic records, the presence of many irrelevant features, and the requirement for a human-labeled training set. In this paper, we describe a knowledge-driven online multimodal automated phenotyping (KOMAP) system that i) generates a list of informative features by an online narrative and codified feature search engine (ONCE) and ii) enables the training of a multimodal phenotyping algorithm based on summary data. Powered by composite knowledge from multiple EHR sources, online article corpora, and a large language model, features selected by ONCE show high concordance with the state-of-the-art AI models (GPT4 and ChatGPT) and encourage large-scale phenotyping by providing a smaller but highly relevant feature set. Validation of the KOMAP system across four healthcare centers suggests that it can generate efficient phenotyping algorithms with robust performance. Compared to other methods requiring patient-level inputs and gold-standard labels, the fully online KOMAP provides a significant opportunity to enable multi-center collaboration.

Evaluating the Bias, type I error and statistical power of the prior Knowledge-Guided integrated likelihood estimation (PIE) for bias reduction in EHR based association studies.

Evaluating the Bias, type I error and statistical power of the prior Knowledge-Guided integrated likelihood estimation (PIE) for bias reduction in EHR based association studies.

Refining chronic pain phenotypes: A comparative analysis of sociodemographic and disease-related determinants using electronic health records.

Refining chronic pain phenotypes: A comparative analysis of sociodemographic and disease-related determinants using electronic health records.

Fostemsavir resistance in clinical context: a narrative review.

Fostemsavir, a prodrug of the first-in-class gp120-directed attachment inhibitor temsavir, is indicated in combination with other antiretrovirals for the treatment of multidrug-resistant HIV-1 in adults who are heavily treatment-experienced (HTE). Temsavir binds to HIV-1 gp120, close to the CD4 binding site, preventing the initial interaction of HIV-1 with CD4 on the host cell. Amino acid substitutions at four positions in gp120 have been identified as important determinants of viral susceptibility to temsavir (S375H/I/M/N/T/Y, M426L/P, M434I/K, M475I), with a fifth position (T202E) recently described. For most currently circulating group M HIV-1 subtypes, the prevalence of these resistance-associated polymorphisms (RAPs) is low. As with many other antiretrovirals, the impact of RAPs is modified by other changes in the target molecule. Different regions of gp120 interact to modify the temsavir binding pocket, with multiple amino acids playing a role in determining susceptibility. Extensive variability of HIV-1 gp120 means the susceptibility of clinical isolates to temsavir is also highly variable. Importantly, in vitro measurement of the susceptibility of clinical isolates to temsavir does not necessarily capture the range of susceptibilities of the heterogeneous mix of viruses generally present in each isolate. Due to these factors and limited phenotypic clinical data, thus far, no relevant phenotypic cutoff or genotypic algorithms have been derived that reliably predict response to fostemsavir-based therapy in individuals who are HTE; therefore, pre-treatment temsavir resistance testing may be of limited benefit. In the phase III BRIGHTE study, re-suppression after virologic failure was observed in some participants despite treatment-emergent genotypic and/or phenotypic evidence of reduced temsavir susceptibility, and substantial CD4+ T-cell count increases occurred even among participants with HIV-1 RNA ⩾40 copies/mL at Week 240. Clinical management of people who are HTE and experience virologic failure during treatment with fostemsavir-based regimens requires an individualized approach with consideration of potential benefits beyond virologic suppression.

Application of an Externally Developed Algorithm to Identify Research Cases and Controls from EHR Data: Trials and Triumphs.

The use of electronic health records (EHRs) in research demands robust and interoperable systems. By linking biorepositories to EHR algorithms, researchers can efficiently identify cases and controls for large observational studies (e.g., genome-wide association studies). This is critical for ensuring efficient and cost-effective research. However, the lack of standardized metadata and algorithms across different EHRs complicates their sharing and application. Our study presents an example of a successful implementation and validation process.This study aimed to implement and validate a rule-based algorithm from a tertiary medical center in Tennessee to classify cases and controls from a research study on rotator cuff tear (RCT) nested within a tertiary medical center in North Texas and to assess the algorithm's performance.We applied a phenotypic algorithm (designed and validated in a tertiary medical center in Tennessee) using EHR data from 492 patients enrolled in a case-control study recruited from a tertiary medical center in North Texas. The algorithm leveraged the international classification of diseases and current procedural terminology codes to identify case and control status for degenerative RCT. A manual review was conducted to compare the algorithm's classification with a previously recorded gold standard documented by clinical researchers.Initially the algorithm identified 398 (80.9%) patients correctly as cases or controls. After fine-tuning and correcting errors in our gold standard dataset, we calculated a sensitivity of 0.94 and a specificity of 0.76. The implementation of the algorithm presented challenges due to the variability in coding practices between medical centers. To enhance performance, we refined the algorithm's data dictionary by incorporating additional codes. The process highlighted the need for meticulous code verification and standardization in multi-center studies.Sharing case-control algorithms boosts EHR research. Our rule-based algorithm improved multi-site patient identification and revealed 12 data entry errors, helping validate our results.

Beyond the loss of beta cells: a quantitative analysis of islet architecture in adults with and without type 1 diabetes

Aims/hypothesisThe organisation and cellular architecture of islets of Langerhans are critical to the physiological regulation of hormone secretion but it is debated whether human islets adhere to the characteristic mantle–core (M-C) structure seen in rodents. It is also unclear whether inherent architectural changes contribute to islet dysfunction in type 1 diabetes, aside from the loss of beta cells. Therefore, we have exploited advances in immunostaining, spatial biology and machine learning to undertake a detailed, systematic analysis of adult human islet architecture in health and type 1 diabetes, by a quantitative analysis of a dataset of >250,000 endocrine cells in >3500 islets from ten individuals.MethodsFormalin-fixed paraffin-embedded pancreatic sections (4 μm) from organ donors without diabetes and living donors with recent-onset type 1 diabetes were stained for all five islet hormones and imaged prior to analysis, which employed a novel automated pipeline using QuPath software, capable of running on a standard laptop. Whole-slide image analysis involved segmentation classifiers, cell detection and phenotyping algorithms to identify islets, specific cell types and their locations as (x,y)-coordinates in regions of interest. Each endocrine cell was categorised into binary variables for cell type (i.e. beta or non-beta) and position (mantle or core). A χ2 test for independence of these properties was performed and the OR was considered to estimate the effect size of the potential association between position and cell type. A quantification of the M-C structure at islet level was performed by computing the probability, r, that the observed number of non-beta cells in the mantle is due to a random arrangement. The distribution of the r values for the islets in the study was contrasted against the r values of a digital population of equivalent randomly arranged islets, termed digital siblings. Both distributions of r values were compared using the earth mover’s distance (EMD), a mathematical tool employed to describe differences in distribution patterns. The EMD was also used to contrast the distribution of islet size and beta cell fraction between type 1 diabetes and control islets.ResultsThe χ2 test supports the existence of a significant (p<0.001) relationship between cell position and type. The effect size was measured via the OR <0.8, showing that non-beta cells are more likely to be found at the mantle (and vice versa). At the islet level, the EMD between the distributions of r values of the observed islets and the digital siblings was emd-1d=0.10951 (0<emd-1d<1). The transport plan showed a substantial group of islets with a small r value, thus supporting the M-C hypothesis. The bidimensional distribution (beta cell fraction vs size) of islets showed a distance emd-2d=0.285 (0<emd-2d<2) between the control and type 1 diabetes islets. The suffixes ‘-1d’ and ‘-2d’ are used to distinguish the comparison between the distribution of one and two variables.Conclusions/interpretationUsing a novel analysis pipeline, statistical evidence supports the existence of an M-C structure in human adult islets, irrespective of type 1 diabetes status. The methods presented in the current study offer potential applications in spatial biology, islet immunopathology, transplantation and organoid research, and developmental research.Graphical

ARCH: Large-scale knowledge graph via aggregated narrative codified health records analysis.

ARCH: Large-scale knowledge graph via aggregated narrative codified health records analysis.

P-679. Risk Factors Shared Among Common Respiratory Viruses

Abstract Background Upper respiratory infections affect patients of all ages, with severity influenced by demographics, comorbidities, and immunosuppression. Studies often determine risk factors associated with severe disease, but rarely are multiple upper respiratory pathogens compared within the same cohort. Using electronic health record (EHR) phenotyping of 267,031 participants with relevant EHR data enrolled in the National Institute of Health’s All of Us Research Program (All of Us), we identified and validated cohorts of participants, and revealed shared risk factors for severity among upper respiratory viruses. Methods We developed EHR phenotype algorithms for eight respiratory viruses using pathogen-specific billing codes, test results, and treatment-dose antivirals. We identified episodes of infection over 90-day periods and validated these using CDC national detection rates over the same period. We then used multivariable logistic regression to compare demographics and identify predictors of hospital admission, adjusting for age, sex, self-identified race and ethnicity, BMI, smoking status, and EHR record length. Results Scaled episode and testing counts paralleled CDC data, with expected seasonal trends observed for influenza, parainfluenza, respiratory syncytial virus (RSV), rhinovirus, human metapneumovirus (hMPV), SARS-CoV-2, and human coronavirus (hCOV) (RSV, hCOV, influenza, and SARS-CoV-2 shown in Figure 1). Across viruses, hospital admission was associated with older age, male sex, underweight, severe obesity, self-identified Black race, and smoking (Table 1). Notably, significant predictors of admission included COPD, asthma, heart failure, ischemic heart disease, hypertension, diabetes, chronic kidney disease, and immunodeficiency, even after adjusting for confounding variables (Figure 2). Conclusion In this study we established and validated an EHR phenotyping algorithm, and demonstrated that clinical risk factors — particularly cardiopulmonary, metabolic, and immunological comorbidities — are associated with severe outcomes among multiple upper respiratory viruses. These results suggest that universal prevention interventions may mitigate the impact of respiratory viral infections in high-risk populations. Disclosures All Authors: No reported disclosures

CohortDiagnostics: Phenotype evaluation across a network of observational data sources using population-level characterization.

This paper introduces a novel framework for evaluating phenotype algorithms (PAs) using the open-source tool, Cohort Diagnostics. The method is based on several diagnostic criteria to evaluate a patient cohort returned by a PA. Diagnostics include estimates of incidence rate, index date entry code breakdown, and prevalence of all observed clinical events prior to, on, and after index date. We test our framework by evaluating one PA for systemic lupus erythematosus (SLE) and two PAs for Alzheimer's disease (AD) across 10 different observational data sources. By utilizing CohortDiagnostics, we found that the population-level characteristics of individuals in the cohort of SLE closely matched the disease's anticipated clinical profile. Specifically, the incidence rate of SLE was consistently higher in occurrence among females. Moreover, expected clinical events like laboratory tests, treatments, and repeated diagnoses were also observed. For AD, although one PA identified considerably fewer patients, absence of notable differences in clinical characteristics between the two cohorts suggested similar specificity. We provide a practical and data-driven approach to evaluate PAs, using two clinical diseases as examples, across a network of OMOP data sources. Cohort Diagnostics can ensure the subjects identified by a specific PA align with those intended for inclusion in a research study. Diagnostics based on large-scale population-level characterization can offer insights into the misclassification errors of PAs.

Standardized patient profile review using large language models for case adjudication in observational research

Using administrative claims and electronic health records for observational studies is common but challenging due to data limitations. Researchers rely on phenotype algorithms, requiring labor-intensive chart reviews for validation. This study investigates whether case adjudication using the previously introduced Knowledge-Enhanced Electronic Profile Review (KEEPER) system with large language models (LLMs) is feasible and could serve as a viable alternative to manual chart review. The task involves adjudicating cases identified by a phenotype algorithm, with KEEPER extracting predefined findings such as symptoms, comorbidities, and treatments from structured data. LLMs then evaluate KEEPER outputs to determine whether a patient truly qualifies as a case. We tested four LLMs including GPT-4, hosted locally to ensure privacy. Using zero-shot prompting and iterative prompt optimization, we found LLM performance, across ten diseases, varied by prompt and model, with sensitivities from 78 to 98% and specificities from 48 to 98%, indicating promise for automating phenotype evaluation.

Toward a Computable Phenotype for Determining Eligibility of Lung Cancer Screening Using Electronic Health Records.

Lung cancer screening (LCS) has the potential to reduce mortality and detect lung cancer at its early stages, but the high false-positive rate associated with low-dose computed tomography (LDCT) for LCS acts as a barrier to its widespread adoption. This study aims to develop computable phenotype (CP) algorithms on the basis of electronic health records (EHRs) to identify individual's eligibility for LCS, thereby enhancing LCS utilization in real-world settings. The study cohort included 5,778 individuals who underwent LDCT for LCS from 2012 to 2022, as recorded in the University of Florida Health Integrated Data Repository. CP rules derived from LCS guidelines were used to identify potential candidates, incorporating both structured EHR and clinical notes analyzed via natural language processing. We then conducted manual reviews of 453 randomly selected charts to refine and validate these rules, assessing CP performance using metrics, for example, F1 score, specificity, and sensitivity. We developed an optimal CP rule that integrates both structured and unstructured data, adhering to the US Preventive Services Task Force 2013 and 2020 guidelines. This rule focuses on age (55-80 years for 2013 and 50-80 years for 2020), smoking status (current, former, and others), and pack-years (≥30 for 2013 and ≥20 for 2020), achieving F1 scores of 0.75 and 0.84 for the respective guidelines. Including unstructured data improved the F1 score performance by up to 9.2% for 2013 and 12.9% for 2020, compared with using structured data alone. Our findings underscore the critical need for improved documentation of smoking information in EHRs, demonstrate the value of artificial intelligence techniques in enhancing CP performance, and confirm the effectiveness of EHR-based CP in identifying LCS-eligible individuals. This supports its potential to aid clinical decision making and optimize patient care.

Electronic Health Records-based identification of newly diagnosed Crohn's Disease cases.

Electronic Health Records-based identification of newly diagnosed Crohn's Disease cases.